AI Security - Netskope Technical Documentation

AI security refers to the protection of artificial intelligence (AI) systems from threats that target their integrity, confidentiality, and availability. As AI becomes more integrated into digital infrastructure, AI cyber security has emerged as a crucial field focused on defending both AI-powered technologies and systems against misuse or attack. To learn more: What is AI Security?

Netskope provides comprehensive AI visibility and control with multi-level services to secure employee access to Gen AI applications.

Access Control

You can use Netskope access controls to:

Identify applications. Netskope can identify 1,800+ Gen AI apps to control usage.
Manage user interactions with activity-based policies.
Configure instance-based policies to define enterprise instances and apply tailored policies.
Leverage application tagging and restrict or permit AI app usage based on defined attributes.
Gain comprehensive visibility into user interactions with AI services via a centralized dashboard that tracks AI SaaS and web traffic, along with detailed usage analytics.

To learn more: Mastering AI Adoption with End-to-end Security, Everywhere.

In the Netskope Cloud, you can leverage AI Security Guardrails, AI Gateway, and AI Red Teaming for AI real-time (inline) security.

AI Gateway

AI Gateway is a software layer designed to act as a secure gateway for intercepting app-to-app traffic between AI agents and AI models, including large language models (LLMs).

In your own deployments (VPC or on-prem), AI Gateway enables applications to access LLMs securely with capabilities such as authentication, rate limiting, prompt/response filtering, monitoring, logging, content moderation, and DLP.

AI Gateway supports the most popular LLM models today:

OpenAI
Google Gemini
Anthropic Claude

With more models coming soon.

It’s available as a virtual appliance that you can deploy in their environment with support for:

Public Cloud: AWS
Private Cloud: VMware ESXi

You can configure and manage AI Gateway—including gateway setup, policy creation, content inspection profiles, logs, and dashboards—all via the Netskope UI.

AI Red Teaming

AI Red Teaming provides vulnerability assessments to secure your LLMs pre-deployment and identify production drifts to meet compliance and ensure quality continuity.

You can deploy the AI Red Teaming service to:

Simulate Adversarial Attacks: Test your LLMs against a wide range of attack scenarios.
Uncover Weaknesses: Pinpoint flaws in your LLM’s alignment, robustness, and content moderation capabilities.
Assess Safety Risks: Proactively identifying potential vulnerabilities, prompt injections, jailbreaking, data leakage, privacy violations, unethical responses, harmful content, and safety gaps.

AI Guardrails

AI Guardrails is an inline AI security module that complements data and threat protection by:

protecting from AI threats like prompt injection and jailbreaking.
enforcing content moderation guardrails for responsible AI usage by detecting and protecting from many categories of AI prompts and responses, such as inappropriate/harmful content, hate, bias, weapons as well as copyright or patented data that expose a new risk vector for organizations.
leveraging a new AI inspection engine that classifies the prompts and responses by mapping them to MITRE ATLAS and OWASP Top 10 LLM frameworks for adversary tactics, techniques, and real-world case studies targeting AI systems.

Agent Guardrails

AI agents require broad access to perform complex tasks autonomously, often without human intervention. However, they can also execute high-impact actions—such as bulk-deleting emails, pushing code, modifying inbox rules, or cloning repositories—which can lead to unintended consequences.

Agent Guardrails gives admins real-time control over these actions through intent-, risk-, and access-based policies. These guardrails apply directly to agent traffic with zero impact on normal user activity.

Agentic Broker

Securing communications and components including identities in agentic implementations has become a new frontier in securing generative AI. With the rapid adoption of the model context protocol and the development of agentic systems, organizations across multiple industries have a need to secure MCP and agentic communications. Let’s take a look at the features launched by Netskope to secure MCP communications.

Model Context Protocol (MCP) is an open standard that enables AI models to connect securely to data sources and tools. Netskope’s Agentic Broker provides administrators with the tools needed to manage the visibility, and access control of MCP implementations, whether they are running locally, in containers, or on remote servers and an inventory of publicly available MCP servers (remote and code repositories).

Risk Framework Mapping

These new AI products map to the OWASP Top10 LLM and MITRE ATLAS frameworks that provide adversary tactics, techniques, and real-world case studies targeting AI systems to empower organizations, security professionals, AI practitioners, and policymakers with comprehensive, actionable guidance to ensure the secure development, deployment, and governance of generative AI systems.

Netskope provides comprehensive AI visibility and control with multi-level services to secure employee access to Gen AI applications.

Access Control

You can use Netskope access controls to:

Identify applications. Netskope can identify 1,800+ Gen AI apps to control usage.
Manage user interactions with activity-based policies.
Configure instance-based policies to define enterprise instances and apply tailored policies.
Leverage application tagging and restrict or permit AI app usage based on defined attributes.
Gain comprehensive visibility into user interactions with AI services via a centralized dashboard that tracks AI SaaS and web traffic, along with detailed usage analytics.

To learn more: Mastering AI Adoption with End-to-end Security, Everywhere.

In the Netskope Cloud, you can leverage AI Security Guardrails, AI Gateway, and AI Red Teaming for AI real-time (inline) security.

AI Gateway

AI Gateway is a software layer designed to act as a secure gateway for intercepting app-to-app traffic between AI agents and AI models, including large language models (LLMs).

AI Gateway supports the most popular LLM models today:

OpenAI
Google Gemini
Anthropic Claude

With more models coming soon.

It’s available as a virtual appliance that you can deploy in their environment with support for:

Public Cloud: AWS
Private Cloud: VMware ESXi

You can configure and manage AI Gateway—including gateway setup, policy creation, content inspection profiles, logs, and dashboards—all via the Netskope UI.

AI Red Teaming

AI Red Teaming provides vulnerability assessments to secure your LLMs pre-deployment and identify production drifts to meet compliance and ensure quality continuity.

You can deploy the AI Red Teaming service to:

Simulate Adversarial Attacks: Test your LLMs against a wide range of attack scenarios.
Uncover Weaknesses: Pinpoint flaws in your LLM’s alignment, robustness, and content moderation capabilities.
Assess Safety Risks: Proactively identifying potential vulnerabilities, prompt injections, jailbreaking, data leakage, privacy violations, unethical responses, harmful content, and safety gaps.

AI Guardrails

AI Guardrails is an inline AI security module that complements data and threat protection by:

protecting from AI threats like prompt injection and jailbreaking.
enforcing content moderation guardrails for responsible AI usage by detecting and protecting from many categories of AI prompts and responses, such as inappropriate/harmful content, hate, bias, weapons as well as copyright or patented data that expose a new risk vector for organizations.
leveraging a new AI inspection engine that classifies the prompts and responses by mapping them to MITRE ATLAS and OWASP Top 10 LLM frameworks for adversary tactics, techniques, and real-world case studies targeting AI systems.